What happened
Lesotho formally approved its National Artificial Intelligence Policy in the second half of 2025 and posted the final policy document on August 17, 2026. The policy establishes a three-tier governance architecture comprising an AI policymaking function, an independent AI regulator, and a multi-stakeholder Data and AI Committee. It also calls for AI-specific legislation, risk controls and audits, public-participation mechanisms, domestic skills and research capacity, computing infrastructure, and a 2025–2030 monitoring framework.
Potential direction of power
Dispersion · Dispersion +1 — Leaning dispersion
The policy is a concrete national institutional action rather than a consultation or announcement. Its dispersion significance comes from explicitly linking AI governance to domestic regulatory, technical, research, and infrastructure capacity—an effort by a small developing state to preserve meaningful national agency rather than remain principally dependent on AI systems and infrastructure controlled elsewhere.
Lesotho provides a useful contestable-scale example: access to AI or even hosting data-center infrastructure does not itself create national AI capacity. The policy treats regulatory institutions, skills, research capability, public participation, and infrastructure sovereignty as complementary elements of effective public power.
Window of contestability
Opens
The policy creates several formal points for outside participation and scrutiny through the multi-stakeholder committee, public-participation mechanisms, audits, reporting, and monitoring indicators. The principal uncertainty now concerns implementation: whether the proposed institutions receive sufficient legal authority, resources, and independence to perform those functions effectively.
